348 ALASKAN GLACIER STUDIES retardation such as this, it is, of course, not safe to estimate the ages of trees by diameter alone. The trees must be cut down and the rings counted. The absence of trees in southern Harriman Fiord, bringing up the possibility of recent expansion of Harriman Glacier, may be due to the fact that conifers are not always seeded at once after a glacier retreats,' while alders and willows may start sooner. Accordingly the length of time since retreating glaciers have uncovered a tract, as determined by vegetation, should always be regarded aa a Tm'nimimri. In the case of the barren zone around Barry Glacier, however, we differ from Grant and Higgins who believe that the advance to its outer limit was probably 25 years or more ago.1 We have direct evidence that seeding did take place there at once, for in the field we first determined, by ages of trees that the last advance was at least 11 years before (1899 or earlier), and subsequently after reaching home, found the records of the army officers who witnessed the activity in May, 1898, exactly 12 years before. This advance ceased during that year, however* and there was considerable melting before June, 1899, so that willows and alders were able to start on the abandoned lateral moraines and in the edges of the barren zone during the summer of 1899. Thus they were eleven years old in 1910. Similar immediate occupation by vegetation was found to have occurred on the elevated beaches in Yakutat Bay, where, as in the Barry Glacier barren zone, the proximity of mature plants gave opportunity for immediate seeding. With these facts and limitations in mind, the following relationships of vegetation to the Barry, Serpentine, Toboggan, and Baker Glaciers in Harriman Fiord may be profitably considered. As already stated, there are conspicuous barren zones around the termini and lower borders of a number of the Prince William Sound glaciers (Pis. CXXXIX and CXXXH, B), as a result of the fact that the glaciers have melted back from their farthest recent advance. There was such a barren zone around Columbia Glacier when Gilbert visited it in 1899, but before 1909 most of this barren zone was overridden by the advancing glacier and by 1910 it was covered, as were most of the barren zones which had bordered the College Fiord glaciers from 1899 to 1909. In Harriman Fiord, on the other hand, these barren zones were still present near most of the glaciers in 1910, Our first impression was that the advances which caused these barren zones were synchronous and that an examination of the ages of shrubs that had sprung up in the barren zones would reveal the approximate date at which the last advance of all the glaciers had taken place. The vegetation near Serpentine Glacier was studied and in its barren zone none of the scattered willows, alders, and spruces were found to be more than 27 years old. An earlier moraine outside this belt had trees up to 93 years of age. Baker Glacier, just south of Serpentine, had a nearly-barren inner moraine with shrubs 18 years old and a more-thickly-f crested outer moraine with spruces up to 110 years old. Toboggan Glacier had 20 year old shrubs on the last moraine of its barren zone and conifers of 70 years growth on an older moraine. Barry Glacier was bordered by a barren zone with the oldest shrubs only 11 years old; and here there are not two belts, but one, outside of which, on each margin of the glacier, a very perfect push-moraine ridge separates the barren zone from the thick forest 100 to 225 years old. We cut down many trees and counted rings in determining these ages, only the oldest ones being cited here. At the i Bull. Amer. Geog. Soc., Vol. XT.TTI, 1911, p. 881.